Help us improve
Share bugs, ideas, or general feedback.
From feature-todo
12-stage feature implementation workflow with Serena MCP, anti-duplication checks, and plugin invocation. Use when implementing new features, starting a multi-stage development workflow, or when you need structured guidance through planning, implementation, testing, and deployment phases. Trigger with "feature implementation", "implement feature", "feature workflow", or "/feature-todo:implement".
npx claudepluginhub omara1-bakri-org/local-plugins --plugin feature-todoHow this skill is triggered — by the user, by Claude, or both
Slash command
/feature-todo:feature-workflowThis skill is limited to the following tools:
The summary Claude sees in its skill listing — used to decide when to auto-load this skill
```bash
Implements features using parallel subagents with scope control, reflection, and MCP servers for memory/context. Activates on implement/build/create requests in JS/TS projects.
Manages a 4-phase feature development pipeline: research, implementation planning, progress tracking, and phased execution for major features. Supports status checks and next-step guidance via subcommands.
Guides technical evaluation of code review feedback: read fully, restate for understanding, verify against codebase, respond with reasoning or pushback before implementing.
Share bugs, ideas, or general feedback.
/feature-todo:implement [name] # Full workflow
/feature-todo:implement [name] --project=X # Custom project
/feature-todo:stage [1-12] [name] # Single stage
/feature-todo:checklist [stage] # View checklist
// ALWAYS activate first
mcp__serena__activate_project('root') // or custom project name
// Core commands
mcp__serena__get_project_overview() // Project structure
mcp__serena__find_symbol('name') // Find class/function
mcp__serena__search_code('query') // Semantic search
mcp__serena__get_type_hierarchy('X') // Find base classes
mcp__serena__find_references('symbol') // Find usages
mcp__serena__get_file_tree() // File structure
BEFORE creating ANY new file:
// 1. Search for existing
mcp__serena__find_symbol('NameYouWantToCreate')
mcp__serena__search_code('what you are building')
// 2. Find base classes
mcp__serena__get_type_hierarchy('PotentialBase')
// 3. Decision
// - If similar exists → EXTEND it
// - If nothing similar → CREATE (follow existing patterns)
Goal: Map codebase, create todos
Invoke:
/task-router [feature]Task subagent_type="Explore"Serena:
get_project_overview()
get_file_tree()
search_code("similar features")
Output: 12 TodoWrite items, documented patterns
Goal: Design aligned with existing patterns
Invoke:
Task subagent_type="architecture-strategist"Task subagent_type="database-architect"Serena:
search_code("architecture patterns")
get_type_hierarchy("BaseService")
Output: Data models, API contracts
Goal: Deep codebase understanding
Invoke:
Task subagent_type="pattern-recognition-specialist"Task subagent_type="Explore" thoroughness="deep"Serena:
find_symbol("Service")
find_references("ExistingCode")
Output: Reusable code list, extend vs create decisions
Goal: Prepare environment
Actions:
git checkout -b feature/[name]
npm install [dependencies]
Output: Feature branch, migrations
Goal: Build WITHOUT duplication
⚠️ CRITICAL: Before EVERY new file:
find_symbol("NameToCreate")
search_code("description")
get_type_hierarchy("BaseClass")
Invoke:
Task subagent_type="backend-architect"Task subagent_type="frontend-developer"Output: Working code extending existing patterns
Goal: >80% coverage
Invoke:
Task subagent_type="test-automator"Task subagent_type="security-sentinel"Serena:
search_code("test helper")
find_symbol("fixture")
Output: All tests passing
Goal: Quality verification
Invoke:
Task subagent_type="code-reviewer"Task subagent_type="code-simplicity-reviewer"Verify: No duplicates created
Output: Clean, reviewed code
Goal: Complete docs
Invoke:
Task subagent_type="docs-architect"Output: JSDoc, README, CLAUDE.md updated
Goal: Ready for deploy
Invoke:
Task subagent_type="deployment-engineer"Output: Docker, CI/CD, monitoring configured
Goal: Final verification
Invoke:
Task subagent_type="deployment-verification-agent"Output: Staging verified, rollback tested
Goal: Production release
Actions:
Output: Production live
Goal: Monitor and document
Invoke:
Task subagent_type="observability-engineer"Output: 24h monitoring, retrospective, changelog
If Serena unavailable:
Grep for text searchGlob for file patternsserena-unavailableIf plugin/agent unavailable:
Feature complete when:
For Ralph loop output:
<promise>FEATURE_COMPLETE</promise>